time at the start of Task 2 was dedicated to obtaining a guayule-based material with the properties required for use as a recycling agent (RA). Ultimately, eleven different
combinations of guayule feedstock material, solvent(s), and extraction/recovery processes were investigated. The general criteria for a more in-depth evaluation of an extract were 1) similarity of the extract’s temperature-viscosity relationship to those of commercially- available, petroleum-based RAs, 2) simplicity and relative safety of the extraction and recovery process, and 3) sufficient yield (weight percent of material recovered from a particular guayule-based feedstock).
The plant or plant-extract precursor materials investigated were three different plant feedstocks, two guayule-based materials resulting from the solvent extraction of one of the plant feedstocks, and the dried latex (bulk rubber). The plant feedstocks were USDA-ARS-supplied chipped whole-shrub (WS), and Yulex-supplied post-latex- extraction (PLE) bagasse (fibrous residue leftover after 90-95% of the high molecular weight rubber was removed), and waste-stream leaf and attached stems. The whole-shrub and waste-stream leaf/stem feedstocks were pulverized prior to solvent extraction using a
horizontal shaft impactor. The PLE bagasse was already in a finely-macerated state. Figure 3.1 shows the guayule-based feedstock supplied by Yulex and USDA-ARS, and used for the greatest portion of the extractions.
The earliest acetone-extraction and recovery techniques of the guayule resin followed procedures outlined in the literature [11] [70]. The USDA-ARS supplied freshly harvested, whole-shrub guayule plants that had been field-dried, then run through a chipper and reduced to a size passing a 3/8 inch screen. The chipped whole-shrub was further reduced in size (-1/16 inch) using the horizontal shaft impactor. The pulverized guayule plant material was then soaked in acetone overnight with occasional stirring. The following morning, the acetone-resin solution was drained off, filtered, and stored in glass containers. A second soaking of the guayule plant material in acetone was performed but did not last as long as the primary soaking. The acetone-resin solution from the second soaking was, again, drained off, filtered, and stored in glass containers. A final wash of the guayule plant material with acetone was performed with a small amount of acetone and this wash solution was also filtered and stored in glass containers. The filtered solution was further clarified using the large centrifuge specified in the Abson binder recovery procedure, AASHTO T 170 [70]. The clarified acetone-resin solution (i.e. miscella) was recovered (i.e. desolventized) using the distillation equipment also specified in AASHTO T 170.
The quantity of recovered guayule resin using the method just described was minimal. Additionally, the Abson method of recovery was known to be somewhat dangerous in that, if not carefully watched, the material in the recovery flask could “bump” or “burp” and be violently discharged from the flask. Because the volumes of
resin required for this study were so large, a 20 liter solvent recovery device (SRD) was, therefore, purchased. Figure 3.2 shows the SRD, the primary distillation device, and the rotary evaporator, the final distillation device, which were used for the remainder of the study. General extraction and recovery details for this scaled-up procedure are given in Appendix B.
Figure 3.1. Guayule Plant Feedstocks for Extractions
Screening the eleven different extracts for further evaluation was sometimes based simply on visual observation. For example, the first acetone-extracted resins from
the chipped whole-shrub and the PLE bagasse would flow when slightly heated, and would “string out” (exhibit some cohesion and ductility) when pulled from the container with a small spatula, but would become stiff and/or brittle (glassy) upon cooling to room temperature. This behavior made these extracts ill-suited as a RA but demonstrated a potential to improve the high-temperature performance of FPMs as stiffening agents.
Figure 3.2. Primary and Final Distillation Stations
Simple acetone-extraction performed on the fibrous feedstocks was not producing a material suitable as a RA. The viscosity of these extracts did not indicate the potential for significantly reducing the viscosity of the RAP/RAS binders upon blending, which is the basic purpose of RAs.
The next step beyond simple acetone-extraction was determining the extent to which “maltene-like” oils and hydrocarbon compounds could be isolated from the
acetone-extracted PLE bagasse resin. Buchanan et al. [71] present a method of
performing acetone-extraction on whole plant material then partitioning the desolventized extractive into oil and polyphenol fractions using hexane (non-polar solvent) and aqueous methanol (polar solvent), respectively. The basic procedure outlined by Buchanan et al. was followed for this study but pentane was used instead of hexane to stay consistent with the standard clay-gel chromatography test for classifying recycling agents, which specifies pentane as the solvent [72]. Although the process outlined by Buchanan et al. involved whole plant material as the beginning feedstock, the PLE bagasse resin was investigated first because the PLE bagasse is a by-product of the rubber extraction process and finding a market for it would be most beneficial to the guayule processing industry.
The procedure, referred to as a liquid-liquid extraction, began with putting 100 grams of the acetone-extracted PLE bagasse resin into a large glass container then incrementally adding several hundred grams of a 90% aqueous methanol solution (90% methanol, 10% water, by weight) to the container while stirring with a glass rod. The aqueous methanol was added until no more resin (more specifically, the solids remaining in the container) would dissolve. At that point, pentane was added in measured
increments while continuing to stir. When the stirring would stop, one could see a definite phase separation with the lighter density, light green colored pentane solution floating on top of the heavier, yellowish-orange colored aqueous methanol solution in the bottom of the container. The pentane addition and stirring continued until all solids had dissolved, except for a few solids that occupied the boundary layer between the two liquid phases. At this point, the majority of the pentane solution was siphoned into large
glass jars and desolventized using the rotary evaporator device. Table 3.1 shows the viscosity test results for the recovered pentane solubles fraction of the acetone-extracted PLE resin.
Table 3.1. Viscosity Comparison for the Pentane Soluble Fraction of the PLE Resin
Temperature (°C) Viscosity (centipoise)
Pentane Solubles Cyclogen L (CycL)
60 1044 438
80 285 93
100 107 31
Table 3.1 shows the viscosity of the pentane solubles in comparison to that of Cyclogen L (CycL), a petroleum-based aromatic oil that is a marketed as a true rejuvenator, and was used throughout this study. The PLE resin pentane solubles are comparable to the CycL in viscosity and are also similar in that CycL is 98.2% pentane soluble (as determined during clay-gel absorption chromatography, to be discussed later). Generating and utilizing the pentane solubles from the guayule materials is the preferred route to obtaining a bio-based, plant-oil-type RA. However, the extraction and recovery process is more complicated, and yield would be considerably smaller on a weight- percent basis of the guayule feedstock. This small investigation yielded approximately 15 grams of pentane solubles from the 100 grams of PLE bagasse resin. This particular PLE resin batch was actually recovered from the small batch of PLE bagasse obtained early in the study. Acetone-extraction was performed on five kilograms of dried PLE bagasse.
Therefore, the pentane-soluble yield using the liquid-liquid extraction procedure was only 0.3% of the dried PLE bagasse feedstock.
The focus returned to single-solvent extractions. The following is a list of the various guayule plant feedstocks and the solvents used for extraction:
• Pulverized whole-shrub: pentane.
• PLE bagasse: toluene, pentane, and hexane.
• Pelletized waste-stream leaf/stem: acetone, to see how it compared to whole- shrub and PLE bagasse acetone-extracts.
• Pulverized waste-stream leaf/stem: hexane.
Because they are essentially waste products (potential co-products) of the latex extraction procedure used by Yulex, the PLE bagasse and the waste-stream leaf/stem materials were the primary feedstocks of interest.
The toluene-extracted resin from the PLE bagasse had viscous properties similar to the acetone-extracted PLE bagasse resin (very stiff), but was slightly less temperature susceptible (i.e. exhibits less change in viscosity for an equal change in temperature). The pentane-extracted resin from the PLE bagasse and the WS possessed viscous properties more appropriate for a recycling agent but pentane is very volatile and dangerous, and expensive. The hexane-extracted resins of both the PLE bagasse and waste-stream leaf/stem materials were similar to the PLE bagasse pentane-extracted resin in terms of viscosity. The hexane-extraction of the PLE bagasse resulted in a 3.0% yield of resin based on the oven-dry (60°C overnight) weight of the bagasse, and the hexane-extraction of the pulverized waste-stream leaf/stem feedstock resulted in a 5.1% yield of resin (LF), also based on the oven-dry (60°C overnight) weight of the leaf/stem feedstock. Because
hexane has the advantage of being less dangerous and less expensive than pentane, hexane became the primary solvent to be further evaluated.
In an effort to improve upon the liquid-liquid extraction procedure discussed earlier, a method used in the “essential oil” production industry was investigated to obtain more of the non-polar compounds present in guayule feedstocks [73]. The basic
methodology begins by extracting the non-aromatic waxes, pigments, and volatile aromatic molecules from the feedstock using a solvent, usually hexane. The solvent is then recovered through distillation leaving a resin (or a “concrete” in essential oil parlance). The waxy materials and “essential oils” in the resin are further separated with an alcohol (e.g. pure ethanol) while warming and stirring the resin/alcohol mixture. The aromatic oils (or the “absolute” in essential oil parlance) and some waxy compounds are dissolved by the alcohol while the majority of the waxy materials remain as a residue. The alcohol solution is then repeatedly cooled and filtered to remove the few dissolved waxy compounds. The alcohol is then recovered through distillation leaving the aromatic or “essential” oils.
For the initial investigation into this method, 50 grams of the hexane-extracted resin (LF) from the pulverized waste-stream leaf/stem feedstock was put in a glass beaker and ~1 liter of pure ethanol was added to the beaker. The beaker was then put on a hot plate and warmed (somewhat below the boiling point of ethanol which is 79°C) while stirring with a glass rod. It could be visually observed that separation was occurring as the ethanol took on a greenish color and the residue in the bottom of the beaker became very sticky. The ethanol/oil solution (miscella) was decanted into a separate container and the residue was covered with another liter of pure ethanol to assure maximum
removal of the oils. This second liter remained clear after a lengthy interval of warming and stirring indicating that the greatest majority of the ethanol-soluble compounds had been removed with the first liter.
The ~2 liter bottle of ethanol miscella was put in a refrigerator at a temperature of ~ 35°F (~ 2°C) overnight then cold-filtered through coffee filters the following day. Several filters (typical mesh size is 10 – 15 micrometers) had to be used because there was a significant amount of waxy material precipitate present in the cooled miscella. The filtered miscella was then put in a freezer at ~ -20°F (~ -30°C) overnight then cold- filtered the next day, removing a smaller, but significant, amount of precipitates. The process of cooling and filtering was repeated one more time using the freezer. The
ethanol was then removed through distillation using a rotary evaporator and the recovered residue was collected for future testing.
Based on the first attempt at partitioning of the “oil” (ethanol-soluble) portion of the LF extract, the yield was 54.4 % of the weight of the LF. Therefore, based on the 5.1% yield of the LF extract from the pulverized leaf/stem feedstock, one would estimate a 2.8% yield of the LF oil based on the weight of the pulverized leaf/stem. This was a big improvement over the 0.3% pentane soluble material yield from the dried PLE bagasse, as described earlier. The ethanol partitioning method was simpler, safer, produced more material, and said material had viscous properties that were more appropriate for use as a RA, relative to the acetone-extracted resins.
Ultimately, however, the ethanol partitioning of the LF extract was not pursued further. The decision was made to focus on the LF material as extracted from the waste- stream leaves and attached stems. The extraction procedure was simple and the yield
averaged 4.4% by weight of the oven-dry (60°C) waste-stream leaf/stem. The LF was dark-green in color, semi-solid at room temperature, and had a piney, pleasant odor. However, the LF also contained some natural rubber which increased the tackiness of any blend in which it was present.
Soon after beginning testing using the LF extract, the Yulex Corporation suggested acetone-extraction on the dried latex (bulk rubber) they produced as a secondary product. The rubber had residual, amber-colored resin in it that 1) was
undesirable for producing a near-white rubber that enables the creation of various grades of rubber, and 2) could be extracted using acetone. Yulex supplied about 50 pounds of the rubber and a method was devised to extract the resin by cutting the rubber into approximately 2 inch cubes, freezing the cubes in liquid nitrogen, pulverizing the frozen cubes, soaking the flaked rubber in acetone for an extended period of time, then
desolventizing the acetone-resin solution in a two-step distillation process. The acetone- extracted rubber resin (RR) was amber in color, flowed at room temperature, and had a slightly pungent, but not unpleasant odor. The acetone-extraction process was relatively simple and the yield averaged about 13% by weight of the rubber. The procedural details are given in Appendix B.
After months of experimentation, the LF and RR were chosen for further evaluation. Each of the two guayule-based materials was compared to an appropriate petroleum-based product of similar viscosity: 1) RR was compared to CycL, and 2) the LF was compared to a PG52-28 binder. PG52-28 is a soft binder often used in FPMs with high contents of RAP and/or RAS, and in circumstances where the contract-specified binder grade is stiffer than PG52-28. Table 3.2 shows a summary of results, in a general
chronological order from top to bottom, of the investigations into obtaining a guayule- based extract that could serve as a RA.
Table 3.2. Summary of Initial Guayule-Based Material Generation Investigations
Plant or Plant Extract
Precursor Materials Solvent Results
Whole-Shrub (WS) Acetone The WS resin is much stiffer and more temperature susceptible than a PG52-28 binder. PLE Bagasse (PLE) Acetone The PLE resin is very stiff. Viscosity-temperature relationship similar to RAP binders. Waste-Stream
Leaf/Stem Acetone
Visual inspection only. Visual observation was similar to acetone-extracted WS resin. Too glassy (brittle) at room temperature.
Acetone-extracted PLE Bagasse Resin
Pentane-Aqueous Methanol Liquid- Liquid Extraction
The extract has good viscous properties comparable to Cyclogen L. No follow-up work: pentane hazard-risk and high cost, and a very complex production method. Very small yield.
Whole-Shrub (WS) Pentane Visual inspection only. The extract demonstrates moderate ductility, high elasticity. No follow-up work: pentane hazard-risk and high cost.
PLE Bagasse (PLE) Pentane Visual inspection only. The extract demonstrates moderate ductility and elasticity. No follow-up work: pentane hazard-risk and high cost.
PLE Bagasse (PLE) Toluene The resin is similar to the acetone-extracted PLE resin in terms of viscosity but is slightly less temperature susceptible.
Waste-Stream
Leaf/Stem Hexane
The extract (LF) has viscosity similar to a PG52-28 but is significantly less temperature susceptible. Simple production method and moderate yield. Demonstrates high ductility, moderate elasticity upon visual inspection. Contains some rubber and is tacky.
PLE Bagasse (PLE) Hexane The extract has a viscosity-temperature relationship similar to the LF material with higher viscosity; i.e. too stiff.
Hexane-extracted LF Cold-filtered Ethanol-partition
The extract is less viscous than a PG46-28 binder but more viscous than Cyclogen L, and more temperature susceptible. Production less complex than pentane- aqueous methanol partitioning. Small yield
Dried Latex (bulk
rubber) Acetone
The rubber resin (RR) has viscosity similar to Cyclogen L but is slightly less temperature susceptible. Relatively simple production method and high yield.